V-I / I-V Converters: Transconductors & TIAs for Sensors
← Back to:Active Filters & Signal Conditioning
V-I / I-V blocks turn sensor current into a measurable voltage (TIA) or force a precise current into a load (transconductor) while keeping the critical node under control. The practical path is always the same: lock the range mapping first, then close the noise and stability budgets under Cin/leakage/protection parasitics, and verify settling, linearity, compliance, and overload recovery with measurable pass criteria.
What is V-I / I-V and where it belongs in the signal chain
V-I (transconductor) and I-V (TIA) are the two “unit converters” that move a system between current domain and voltage domain. The practical goal is not only conversion, but predictable linearity, accuracy, noise, and stability under real sensors, cables, and loads.
A) Definitions & scope boundary (fixed ownership)
- I-V (TIA): converts sensor current to voltage while keeping the input node voltage nearly constant (virtual ground / constant potential behavior).
- V-I (Transconductor): converts an input voltage (or command) to output current that stays nearly independent of load (high output resistance + compliance margin).
- In-scope: linearity, DC accuracy, noise, stability, layout/leakage control, protection impact, and electrochemistry front-end mapping at the signal-conditioning level.
- Out-of-scope: charge amplifier reset details, protection standards catalogs, lock-in demod theory, RMS/log blocks, and ADC sampling theory (only interface-level requirements appear).
B) Typical inputs/outputs (organized by physical constraint)
- Photodiode and other junction-based sensors (node voltage affects capacitance and linearity)
- Ion/current-mode sensors from pA to mA (leakage and bias dominate at the low end)
- Electrochemistry working-electrode current measurement (map electrode behavior into equivalent error terms)
- Constant-current excitation (RTD/bridge sensors; watch compliance + self-heating)
- 4–20 mA or other current-loop style interfaces (load range + protection recovery)
- Electrode drive as part of a larger control loop (this page stays at the conversion block boundary)
C) One-line discriminator (turn concepts into engineering decisions)
- Choose I-V (TIA) when input-node voltage movement changes sensor behavior (capacitance, polarization, or nonlinearity). The converter must “own” the node.
- Choose V-I (Transconductor) when load variation is unavoidable but current must remain controlled. The converter must “own” compliance margin and output resistance.
The left path emphasizes holding the input node (sensor integrity and stability). The right path emphasizes output compliance and load independence.
Topology taxonomy and the decision flow (choose the right core)
Topology choice should be driven by constraints that can be measured: current range, sensor/cable capacitance, bandwidth, allowable input-node movement, and output compliance or load span. The goal is to select a core that keeps the dominant error mechanisms controllable on the bench and stable on the PCB.
A) Taxonomy (families + boundaries)
- Resistive-feedback TIA: simplest core; stability depends strongly on total input capacitance.
- T-network gain extension: expands effective transimpedance without extreme resistor values; verify noise/leakage impact.
- Cf compensation: stabilizes the noise-gain rise created by Cin; trades bandwidth and settling time.
- Integrator-like variants: used when the measurement is inherently slow or charge-related; keep scope at signal-conditioning level.
- Rsense + op-amp: most predictable accuracy; compliance and output swing must be budgeted.
- OTA / gm cell: convenient tuning; verify linearity, temp drift, and output resistance over load.
- Howland-style: can work within strict boundaries; often fails due to resistor mismatch, common-mode limits, or capacitive loads.
B) Decision variables (treat as an input form)
- Range: pA / nA / µA / mA (also note worst-case and transient peaks)
- Bandwidth: DC–LF / kHz / >100 kHz (define required settling and group delay tolerance)
- Cin + cable: sensor capacitance + ESD/protection + routing + connector parasitics
- Input-node constraint: must the input node voltage stay nearly constant? (Yes/No)
- Compliance/load: Vmin..Vmax available, RL range, and any capacitive load behavior
- High Cin + bandwidth → prioritize a TIA core with a clear compensation knob (Cf) and a stability test plan.
- pA/nA regime → prioritize leakage control and bias-current characterization before chasing gain accuracy.
- Wide RL span in V-I → prioritize compliance margin and output swing/drive verification over ideal output resistance assumptions.
C) Decision flow outcomes (recommended core + common failure mode)
Prefer resistive-feedback TIA with explicit Cf compensation sized against total Cin.
Common failure mode: oscillation appears only after connecting the sensor/cable because Cin increases noise gain and removes phase margin.
Prefer Rsense + op-amp transconductor with a clearly budgeted compliance window.
Common failure mode: output current drifts with load because compliance margin is insufficient or output swing collapses under real RL/Cload.
Consider T-network TIA to avoid impractically large Rf, then re-check noise and leakage dominance.
Common failure mode: leakage or bias-current induced offset becomes the dominant term and masks the intended resolution gain.
Use the flow to lock down constraints first. Then choose the simplest core that offers a clear compensation knob and an unambiguous bench pass/fail test.
Core transfer functions: from current to voltage (and voltage to current)
This section defines the minimal small-signal backbone for I-V (TIA) and V-I (transconductor) blocks. Every later discussion on accuracy, noise, stability, and layout maps back to the same model elements and injection points.
A) Ideal transfer (with units)
Vout ≈ − Iin · Rf (A · Ω = V)
The feedback network converts input current into an output voltage through the transimpedance Rf.
Iout ≈ Vin / Rsense (V / Ω = A)
or Iout ≈ Gm · Vin (S · V = A)
The control variable is Vin; the controlled output is current. Output resistance and compliance define real behavior under load.
B) What “virtual ground” and “output resistance” mean (conditions)
- Loop gain must be high in-band (finite Aol and GBW set the error floor).
- Total input capacitance Cin must not consume phase margin (sensor + cable + ESD + routing).
- Output must stay within swing/drive limits; saturation breaks node control and creates slow recovery.
- High Rout keeps Iout stable when RL changes (ideal current source behavior).
- Compliance window must cover worst-case load voltage (limited by supply, swing, and output stage).
- Capacitive loads often require explicit stability planning (do not assume “DC accuracy” implies stability).
C) Non-ideal placeholders (where later budgets connect)
- Aol(s), GBW: finite loop gain adds gain/phase error and sets stability margin.
- Vos, Ib: become DC offset terms (Ib·Rf is often dominant in high-Rf TIAs).
- Cin: raises noise gain at high frequency and can trigger ringing/oscillation.
- Ro + swing/drive limits: collapse compliance or node control under real loads and transients.
The block view is intentionally minimal: each later budget term attaches to one injection arrow or one non-ideal block.
Accuracy budget: where DC errors really come from
DC accuracy is only predictable when each error term is expressed in a consistent reference domain and has a bench measurement hook. Typical datasheet numbers are not sufficient without worst-case conditions, temperature, and board-level leakage control.
A) Budget framework (pick one reference domain)
- TIA: express DC error as equivalent input current error (Ierr) or output voltage error (Verr = Ierr · Rf). Keep units consistent.
- V-I: express DC error as output current error (Ierr), then verify against load span and compliance limits.
- Budget is actionable only if each term links to: measurement point → mitigation action → pass/fail criterion.
B) Gain-related errors (ratio terms that scale with signal)
- Rf tolerance & tempco (TIA): directly sets transimpedance gain error.
- Rsense tolerance & tempco (V-I): sets commanded current error; drift is often dominated by self-heating.
- Reference errors (if a reference is in the loop): map reference drift into Ierr or Verr consistently.
- Sweep a known current/command and fit gain at two points under the same thermal condition.
- Thermal A/B: compare gain after warm-up vs steady-state; watch Rsense self-heating signatures.
C) Offset-related errors (terms that dominate near zero)
- Vos: appears as output offset; verify under the same input bias and temperature conditions.
- Ib · Rf (TIA): often dominates at high Rf; typical Ib is not a guarantee across temperature and humidity.
- Board leakage: contamination and humidity create hidden current paths that look like sensor current.
- Disconnect sensor, short/guard the input node, and observe baseline drift versus environment changes.
- Replace the sensor with a known current source; compare drift to separate instrument drift from sensor drift.
- If touching cables changes baseline, suspect leakage or parasitic coupling before “gain error” assumptions.
D) Temperature/time effects + electrochemistry mapping (modeled as budget terms)
- Drift: specify drift as a slope (per °C or per hour) and measure after a defined warm-up and settling protocol.
- Memory / absorption: treat long-tail recovery as a time-domain error term; quantify the return-to-baseline time constant.
- Electrochemistry: represent electrode polarization/bias behavior as an equivalent input current offset term Ioffset,electrode(T, time, bias) and validate with controlled baseline runs.
A budget is trustworthy only when each segment has a repeatable measurement hook and a documented worst-case condition set.
Noise budget: en/in, resistor noise, and source impedance interaction
A practical noise budget must use a single reference domain, a declared integration bandwidth, and measurement hooks. The goal is a closed loop: calculate → measure → adjust Rf/Cf and amplifier choice without chasing “typical” curves.
A) Reference domain + bandwidth definitions (make numbers comparable)
- Preferred reference: for TIA, express noise as equivalent input current density (A/√Hz), then convert to output via Vn,out ≈ In,eq · Rf.
- Two bandwidths matter: 0.1–10 Hz (low-frequency stability) and an RMS band (system SNR / detection).
- Measurement conditions must be fixed: temperature, warm-up time, input state (shorted / sensor / dummy Cin), shielding/grounding.
- Noise density: en (V/√Hz), in (A/√Hz), and Rf thermal contribution.
- Integrated noise: 0.1–10 Hz and RMS-band value (state the band explicitly).
- A short note on dominant term under the tested condition set.
B) The big three + source noise (who dominates depends on impedance and band)
- Amplifier voltage noise (en): rises with noise gain; Cin and compensation shape how en appears at the output.
- Amplifier current noise (in): turns into voltage noise across impedances; often critical with large Rf and low-frequency work.
- Rf thermal noise: sets a floor; becomes dominant when the amplifier is already low-noise and bandwidth is wide.
- Source noise: sensor physics can dominate (e.g., photodiode shot noise, electrochem baseline noise); model it as an input current noise term.
- Short the input node (with guarding as needed) to reveal front-end baseline noise without sensor contribution.
- Swap Rf by a known ratio: if integrated noise scales strongly with √Rf or Rf, the resistor term is leading.
- Add a known Cin (or cable) and re-measure: a large change implies noise gain / stability coupling is driving the result.
C) 1/f corner and “which noise to report”
- Low-frequency stability: report 0.1–10 Hz to capture drift-like behavior and 1/f dominance.
- Detection/SNR: report integrated RMS noise in the system band (state the band and filtering).
- Time-domain hook: long record → fixed-window calculation → compare before/after warm-up and environment changes.
D) Closed-loop Rf selection (calculate → measure → refine)
- Set swing headroom: choose Rf so that Vout ≈ Iin,max · Rf stays inside the usable output window (reserve margin for recovery).
- Declare the noise band: use the same integration band for computation and measurement (0.1–10 Hz or RMS band).
- Budget in one domain: treat noise sources as injections and combine them consistently before converting to Vout.
- Apply dominance rules: if in·Z or leakage dominates, increasing Rf can worsen baseline; if Rf thermal dominates, bandwidth reduction or smaller Rf is more effective.
- Validate with A/B tests: (a) input shorted, (b) known dummy Cin, (c) real sensor. Differences identify which injection path is leading.
The diagram highlights injection locations and a typical dominance pattern. Final dominance must be verified under the target impedance, Cin, and bandwidth.
Linearity & dynamic range: compliance, swing, and distortion traps
Linearity is governed by a verifiable operating window: input-node control (TIA), compliance (V-I), and large-signal limits such as swing, drive, slew rate, and recovery. Define the window first, then measure for traps at the edges.
A) Dynamic range windows (define constraints before “THD”)
- TIA window: Vout ≈ Iin · Rf must remain inside the usable output swing. When swing collapses, the input node is no longer controlled.
- V-I window: required load voltage Vload ≈ Iout · RL must stay inside compliance. Outside compliance, current is no longer constant.
- Edge behavior matters: distortion and recovery often appear before “hard clipping” due to output stage limits and node-control loss.
B) TIA linearity trap: input node deviation drives sensor nonlinearity
- Mechanism: when Vx deviates from its intended bias, sensor behavior changes (junction capacitance and operating point).
- Symptoms: gain changes with signal level, waveform-dependent settling, and slow return-to-baseline after overload.
- Fast check: observe Vx and Vout near full-scale; increasing swing headroom should reduce the symptom if node control is the root cause.
- Sweep Iin amplitude and compare linearity with two different headroom settings.
- Apply a step overload and measure return-to-baseline time (recovery signature).
C) V-I linearity trap: compliance collapse and finite output resistance
- Compliance collapse: as Vload approaches the compliance boundary, current droops or distorts before obvious clipping.
- Finite Rout: even inside compliance, Iout can vary with RL; the slope directly translates to DC error and distortion under dynamic loads.
- Fast check: sweep RL (or Vload) at fixed command and measure Iout vs Vload to map the usable window.
- Sudden waveform shape change as load varies (window boundary signature).
- Overload causes long recovery due to output stage saturation or current limiting.
- Capacitive loads amplify ringing and distortion if stability is not planned.
D) Large-signal traps (what happens + fast diagnosis)
- Slew limiting: output slope clamps at a constant value; distortion grows rapidly with amplitude and frequency.
- Drive limit / heavy load: distortion increases when load is connected; A/B with a lighter load isolates the cause.
- Crossover artifacts: kinks near zero-crossing on small signals indicate output stage nonlinearity.
- Saturation / clamp recovery: baseline returns slowly after overload; window headroom and limiting strategy must be reviewed.
Define the usable window with margins, then test behavior near edges. Most distortion and recovery traps appear before hard clipping.
Stability & compensation: Cin, Cf, phase margin, and step settling
TIA stability is set by the combined network of sensor/cable/protection capacitance and the feedback path. Compensation must be chosen with a defined acceptance criterion: ringing, settling-to-X%, and repeatability under worst-case Cin.
A) Cin inventory and ownership (what pushes a TIA toward oscillation)
- Sensor capacitance: junction C and bias dependence (often the largest and most variable term).
- Cable/connector capacitance: long leads and connectors shift the stability corner earlier than expected.
- Protection capacitance: ESD/TVS adds a non-zero Cin that must be budgeted.
- Amplifier input capacitance: input C and package parasitics interact with Rf/Cf.
- PCB parasitics: pads, vias, contamination film, and nearby copper raise effective Cin at the high-Z node.
Larger Cin makes noise gain rise earlier, consuming phase margin and increasing ringing sensitivity. Budget Cin as a worst-case stack, not a “typical” number.
B) Cf role and tradeoffs (stability insurance with a price)
- Role: limits the noise-gain rise caused by Cin and improves stability margin.
- Benefit: reduces overshoot and ringing; improves robustness to Cin variation.
- Cost: too large Cf reduces bandwidth and increases settling time; step response becomes slower.
- Risk: too small Cf leaves the loop sensitive; ringing can appear only under certain Cin or probe conditions.
- Worst-case Cin used for validation (sensor + cable + protection + board).
- Target bandwidth and required settling-to-X% time.
- Acceptance thresholds for overshoot and ringing decay.
C) How to choose Cf (engineering loop, not guesswork)
- Stack Cin,max: include sensor/cable/protection/PCB worst-case.
- Define targets: required bandwidth and settling-to-X% time.
- Set a safe noise-gain shape: keep the Cin-driven rise from encroaching the stability margin.
- Validate with steps: overshoot, ringing frequency, and decay speed under Cin,max.
- Freeze acceptance criteria: convert waveforms into pass/fail thresholds for production and field repeatability.
D) Acceptance criteria (ringing, phase-margin clues, settling-to-X%)
- Ringing check: sustained oscillation or slow decay indicates insufficient margin (verify under Cin,max and probing conditions).
- Phase-margin clue (oscilloscope): larger overshoot and multiple visible cycles typically imply reduced margin; use consistent step amplitude and load.
- Settling definition: time to enter and remain within an error band (e.g., X%) after a defined input step.
- Overshoot < __%
- Ringing decays within __ cycles (or within __ ms)
- Settling to __% within __ ms under Cin,max
The plot is a shape guide: Cin drives the rise in noise gain; Cf flattens it. The only acceptable result is a repeatable pass under worst-case Cin using defined step/settling criteria.
Layout & leakage control: guarding, shielding, and parasitics ownership
pA-level failures are rarely mysterious: leakage and parasitics form unintended current paths and capacitance at the high-impedance node. The fix is ownership, a short checklist, and A/B tests that isolate board leakage from sensor behavior.
A) High-Z node rules (short, clean, isolated)
- Shortest path: minimize copper length, pad count, and via count for the input node and Rf/Cf loop.
- Keepout: keep the node away from clocks, fast digital edges, and large copper pours that add parasitic C.
- Isolation: avoid placing resistive dividers, test points, or connectors close to the node unless leakage is controlled.
- Thermal discipline: avoid heat sources that create gradients and humidity-driven film changes around the node.
B) Guard ring and driven guard (when it helps, when it backfires)
- High-Z nodes where surface leakage dominates under humidity and contamination.
- Connectors/cables that create large and variable leakage paths.
- Boards requiring repeatable pA bias stability across environment.
- Driven guard injects noise if the driver is noisy or unstable.
- Guard geometry increases parasitic C and can worsen stability.
- Wrong reference potential creates a new leakage bias path.
C) Fast leakage localization (disconnect / short / humidity A-B)
- Disconnect the sensor: isolate board leakage from source physics.
- Short the input to its reference: observe baseline drift and offset under a defined condition.
- Humidity/temperature A-B: compare drift slope changes; strong correlation indicates film leakage.
- Clean/guard A-B: re-test after cleaning and after enabling/disabling guarding to confirm the path.
D) Routing symmetry and return paths (parasitics are part of the circuit)
- Rf/Cf loop tight: place Rf and Cf adjacent to the amplifier pins; minimize loop area.
- Return continuity: avoid splits and forced detours near sensitive nodes; prevent coupling from high dV/dt returns.
- Protection placement: locate ESD near the connector, but avoid dumping extra C directly onto the high-Z node without budgeting it.
- Checklist fields: input keepout, guard continuity, connector leakage isolation, and parasitic-C budget signoff.
Treat leakage and parasitic capacitance as circuit elements. Map each path (surface/dielectric/connector) to an A/B test, then lock down guard/clean/keepout actions with repeatable pass criteria.
Protection and overload recovery without killing accuracy
Input protection is necessary, but every protection element is also a parasitic capacitor, a leakage source, and a non-linear clamp. The goal is to keep stability, offset, noise, and overload recovery inside measurable acceptance limits.
A) Protection elements and the three hidden penalties (Cpar, Ileak, clamp action)
- Helps: limits clamp injection current and protects the amplifier input.
- Hurts: adds thermal noise and can slow settling if combined with input capacitance.
- Owner: choose by worst-case overload current and allowable bandwidth/settling.
- Helps: prevents deep saturation by limiting input excursions.
- Hurts: leakage (Ileak) becomes an offset term; injected charge can create recovery tails.
- Owner: place and reference carefully; validate across temperature and humidity.
- Helps: absorbs energy at the interface.
- Hurts: parasitic capacitance (Cpar) shifts stability; leakage drifts offset.
- Owner: keep high-C parts away from the high-Z node; isolate with structure.
B) Cparasitic impact: stability and settling penalties that look “random”
- Cpar increases Cin,max: noise gain rises earlier, phase margin shrinks, and ringing becomes probe-sensitive.
- Observed symptoms: added TVS changes ringing frequency/decay; touching the cable or changing probes alters stability.
- Engineering action: move high-C clamps to the interface side and isolate the high-Z node so the effective capacitance is budgeted and repeatable.
Stability must pass under Cin,max with the intended probe and cable configuration. If stability changes with probing, the node is not controlled.
C) Overload recovery: deep saturation, clamp injection, and staged limiting
- Mechanism: the amplifier rails, internal nodes store charge.
- Symptom: long tail before returning to baseline after release.
- Fix pattern: avoid hard-rail hits; apply soft or staged clamping.
- Mechanism: clamp paths inject charge into high-Z nodes.
- Symptom: rebound offset or “stiction” after release.
- Fix pattern: limit injection current with series R and staged limiting.
- Mechanism: protection C and leakage create slow discharge paths.
- Symptom: baseline drifts even when the overload is gone.
- Fix pattern: separate “energy handling” from the high-Z measurement node.
D) Acceptance criteria and measurement hooks (recovery must be measurable)
- Overload amplitude and duration
- Release method (disconnect / return to zero / short)
- Cable and probe configuration (fixed)
- Baseline returns within ±X in T seconds after release
- No secondary drift trend inside T (tail eliminated)
- Same result across temperature and humidity corners
Budget protection parasitics as part of Cin,max and offset. Overload recovery must be validated with a defined stimulus and a pass band: baseline returns within ±X in T seconds after release.
Electrochemistry front-end patterns (scoped to signal conditioning)
Electrochemistry interfaces map cleanly to V-I and I-V building blocks. The scope here is the signal-conditioning layer: where current is converted to voltage, where voltage is converted to current, and where leakage and pickup become measurable error terms.
A) 3-electrode roles (WE / RE / CE) as signal-chain nodes
- WE (Working): where the measurable current flows; this is the I-V (TIA) measurement leg.
- RE (Reference): high-impedance voltage-sense point; accuracy lives here for potential readback.
- CE (Counter): driven electrode; this is the V-I actuation leg that supplies current to achieve the target condition.
- RE sense: leakage and pickup sensitivity
- WE current: TIA noise and offset sensitivity
- CE drive: compliance and overload recovery
B) Two common modes mapped to the V-I / I-V building blocks
- Potentiostatic measurement (measure current): RE is sensed as a voltage node; WE current flows into the TIA and is digitized.
- Galvanostatic excitation (source current): a command voltage is translated to current by a V-I block; compliance determines whether current stays constant under load changes.
- Scope constraint: control-loop tuning is out of scope; only the transduction blocks and error mapping are addressed here.
C) Error mapping: turn electrode and cable effects into budget terms
- Budget term: equivalent current offset (Ieq) or equivalent voltage offset (Veq).
- Hook: compare dummy-cell baseline vs real-cell baseline under the same wiring.
- Budget term: Ileak into the high-Z sense and input nodes (offset and drift).
- Hook: humidity A/B and disconnect/short tests isolate board leakage.
- Budget term: injected voltage noise at RE, injected current noise at WE.
- Hook: cable routing A/B and shield termination A/B (keep conditions fixed).
D) “Where accuracy lives”: partition responsibilities for signoff
- RE sense chain: high impedance, leakage-sensitive; guarding and keepout dominate repeatability.
- WE current chain (TIA): noise/offset dominate; stability must pass with cable and protection installed.
- Drive chain (V-I): compliance and limiting dominate; overload recovery acceptance must be defined.
- Dummy-cell baseline Ileak at humidity corner
- 0.1–10 Hz drift after warm-up (fixed cable)
- Overload recovery: returns within ±X in T seconds
The signal-conditioning scope is explicit: V-I drives CE with compliance constraints, RE is a high-Z sense node, and WE current is converted by a TIA. Leakage and pickup become budget terms that must be validated under cable and environment corners.
Engineering checklist (design + verification + production hooks)
This checklist turns V-I / I-V design into a closed loop: define budgets, verify under worst-case parasitics, and ship with measurable production hooks. Every item below is written as an ownership field plus a pass criterion.
A) Design review checklist (budget ownership before layout freeze)
- Field: sensor + cable + connector + ESD/TVS + PCB node estimate
- Pass: stability and settling verified at Cin,max with intended probing and cabling
- Field: Rf tolerance/tempco, Cf value range and placement constraints
- Pass: ringing and settling-to-X% meet the target at worst-case current and Cin,max
- Field: keepout, guard strategy, cleaning/handling process, connector selection
- Pass: offset/drift remains within budget under humidity and contamination corners
- Field: worst-case load range, output swing limits, saturation avoidance plan
- Pass: no hard-rail saturation in normal operation; recovery spec defined for overload cases
- Field: Cpar impacts stability; Ileak impacts offset/drift; clamp action impacts recovery
- Pass: protection installed still meets PM/settling and offset/drift budgets across corners
B) Bench verification checklist (repeatable tests with fixed fixtures)
- Shorted input baseline (amplifier + PCB floor)
- Open input sensitivity (leakage + pickup exposure)
- Substitute current source / dummy load (system chain)
- Pass: in-band RMS noise & 0.1–10 Hz metric below budget placeholder X
- Use a fixed stimulus and fixed probing/cable setup
- Repeat at Cin,max configuration (sensor/cable/protection)
- Pass: overshoot & ringing decay meet settling-to-X% time target
- Dry vs humid corner; cleaned vs uncleaned board A/B
- Disconnect/short test isolates sensor vs board leakage
- Pass: offset/drift slope stays within budget placeholder X
C) Production hooks (calibration, self-test, reject thresholds)
- Choose points away from rails and clamp activation
- Record conditions: temperature, warm-up time, fixture impedance
- Pass: coefficients remain stable across intended corners
- Relay/digipot paths enable short, inject, and bypass modes
- One-touch isolation: sensor vs board vs ADC chain
- Pass: self-test results predict bench failures with low false rejects
- Noise: in-band RMS and/or 0.1–10 Hz metric < X
- Drift: warm-up drift slope < X over a defined window
- Recovery: overload release returns within ±X in T seconds
Freeze Cin,max and protection parasitics early, then validate stability and settling under the same fixtures used for production signoff. Use explicit thresholds (X, T) derived from the system error budget.
Applications (kept vertical, not a catalog)
Each application below follows the same template: signal chain, primary constraint, quick check, and a measurable pass criterion. The goal is to keep scope vertical and avoid turning this section into a parts catalog.
Photodiode TIA
Chain: photodiode → current → TIA (I-V) → ADC
Primary constraint: Cin,max and sensor linearity sensitivity to node voltage.
Quick check: vary cable length/ESD configuration and confirm ringing/settling changes as expected.
Pass: stability and settling-to-X% meet target at Cin,max and worst-case current.
pA/nA sensing (ultra-high impedance)
Chain: pA/nA source → high-Z node → TIA (I-V) → ADC
Primary constraint: leakage (board, connector, humidity) dominates offset and drift.
Quick check: disconnect sensor, short input mode, and run humidity A/B to isolate Ileak.
Pass: baseline offset/drift remains within placeholder X across humidity corners.
Electrochemistry current measurement
Chain: WE current → TIA (I-V) → ADC, with RE as a high-Z sense node.
Primary constraint: slow drift mapped as equivalent offset terms (Ieq/Veq) plus leakage.
Quick check: compare dummy-cell baseline vs real-cell baseline under identical wiring.
Pass: the mapped offset term stays within placeholder X over the defined measurement window.
RTD / bridge constant-current excitation (V-I)
Chain: DAC/REF → V-I block → RTD/bridge → sense ADC
Primary constraint: compliance and self-heating (current accuracy vs thermal error).
Quick check: sweep load and wiring resistance and confirm current regulation across the range.
Pass: current error < X and thermal-induced error remains within X under the defined excitation.
4–20 mA interface (V-I)
Chain: command voltage → V-I driver → long cable/load → sense/monitor
Primary constraint: load variation plus protection-induced recovery behavior.
Quick check: apply overload/short events and measure baseline return after release with protection installed.
Pass: recovery returns within ±X in T seconds and regulation error remains within X across loads.
Keep applications vertical: describe the chain, pick one primary constraint, validate with one quick check, and sign off with one measurable pass criterion (X/T placeholders come from the system budget).
IC selection logic (TIA / Transconductor): fields + decision order
This section is intentionally scoped to selection fields and a repeatable decision order. It is not a catalog. Use the checklist to compare parts under the same conditions (Cin,max, protection parasitics, load, bandwidth, temperature, and overload policy).
A) Key datasheet fields (grouped by failure mode)
- en / in: only comparable when the vendor provides the exact test setup (source impedance, bandwidth, filtering, and input condition).
- 1/f corner: determines whether low-frequency readings are limited by drift-like noise rather than RMS band noise.
- Input current noise: becomes dominant with large Rf / high-impedance nodes and sets the practical floor for pA/nA sensing.
- Noise vs frequency plots: required to integrate noise into the target bandwidth and to validate 0.1–10 Hz claims.
- GBW / phase behavior: indicates how much Cin,max can be tolerated with a realistic Cf window.
- Input capacitance: must be included in the Cin,max budget (sensor + cable + ESD + op-amp input).
- Overload recovery: selection-critical for optical bursts, electrochem steps, and any clamp event.
- Capacitive-load stability guidance: needed when driving ADC inputs, cables, or protection capacitance.
- Vos + drift: defines baseline offset and temperature slope (must be evaluated with the intended Rf).
- Ib / input leakage: a common “one-line spec” that fails in real humidity/contamination unless the vendor provides conditions.
- ESD / protection structure: can dominate leakage and Cin; the structure matters more than the typical number.
- Output swing vs load: evaluate with the real ADC input, cable, or clamp load (no-load swing is not enough).
- Drive capability: prevents settling errors that look like “noise” or “linearity” problems.
- Compliance (V-I): the available headroom at the output node that keeps current constant over load range.
- Package / cleanliness: high-impedance sensing needs packages and layouts that allow repeatable cleaning and guarding.
B) Decision order (prevents spec cherry-picking)
-
Set Rf (or Rsense / Gm) from range and headroom
Define Imax and the allowed Vout window (or compliance window for V-I). Decide the overload policy (clamp allowed or not). -
Noise feasibility check (en / in / R thermal + source)
Validate under the same bandwidth definition (RMS band or 0.1–10 Hz). Require vendor test conditions to avoid false comparisons. -
Stability under Cin,max + protection parasitics
Confirm a realistic Cf window exists for the sensor+cable+ESD capacitance and still meets the settling target to X% within T. -
Output swing, load, and overload recovery
Check swing-vs-load curves and recovery time after saturation/clamp. If recovery breaks the system, the part is rejected regardless of “typ” noise.
C) Vendor must-answer questions (copy/paste into RFQ)
- Noise conditions: source impedance, bandwidth, filtering, and whether input is shorted/biased (provide the full setup).
- Input leakage vs temperature/humidity: curves or worst-case limits (include board cleanliness assumptions if any).
- Overload recovery: define the stimulus, clamp condition, and the time to return to ±X of baseline.
- Output swing vs load: curves for VOH/VOL or linear range under representative loads and supplies.
- Input capacitance breakdown: typical/max and how ESD structures contribute to Cin and leakage.
- Capacitive load stability guidance: isolation resistor ranges, limits, and validation method.
- Protection leakage: any recommended clamp/TVS types for high-impedance nodes and their leakage characterization method.
- Long-term drift notes: aging behavior or recommended soak/warm-up for production tests.
D) Reference part numbers (starting points only)
These are reference examples to speed up datasheet lookup and lab validation. Final selection must be driven by the decision order above and verified with Cin,max + protection parasitics + real load + temperature.
Request a Quote
FAQs (TIA / V-I): troubleshooting with measurable pass criteria
Scope is strict: oscillation, noise, leakage/drift, linearity, compliance, overload recovery, and measurement traps for TIA/transconductors. Each answer uses a fixed 4-line structure: Likely cause / Quick check / Fix / Pass criteria.
